Nitrate Chemodenitrification by Iron Sulfides to Ammonium under Mild Conditions and Transformation Mechanism

Autotrophic denitrification utilizing iron sulfides as electron donors has been well studied, but the occurrence and mechanism of abiotic nitrate (NO3–) chemodenitrification by iron sulfides have not yet been thoroughly investigated. In this study, NO3– chemodenitrification by three types of iron sulfides (FeS, FeS2, and pyrrhotite) at pH 6.37 and ambient temperature of 30 °C was investigated. FeS chemically reduced NO3– to ammonium (NH4+), with a high reduction efficiency of 97.5% and NH4+ formation selectivity of 82.6%, but FeS2 and pyrrhotite did not reduce NO3– abiotically. Electrochemical Tafel characterization confirmed that the electron release rate from FeS was higher than that from FeS2 and pyrrhotite. Quenching experiments and density functional theory calculations further elucidated the heterogeneous chemodenitrification mechanism of NO3– by FeS. Fe(II) on the FeS surface was the primary site for NO3– reduction. FeS possessing sulfur vacancies can selectively adsorb oxygen atoms from NO3– and water molecules and promote water dissociation to form adsorbed hydrogen, thereby forming NH4+. Collectively, these findings suggest that the NO3– chemodenitrification by iron sulfides cannot be ignored, which has great implications for the nitrogen, sulfur, and iron cycles in soil and water ecosystems.


■ INTRODUCTION
In the natural nitrogen cycle, nitrate (NO 3 − ), nitrite (NO 2 − ), and ammonium (NH 4 + ) are the main inorganic nitrogen species in water. 1 Due to the extensive use of fertilizers in agriculture, along with the discharge of municipal sewage and industrial wastewater, high NO 3 − concentrations in surface and underground water have become a serious environmental issue.This pollution can cause severe environmental damages (i.e., eutrophication) and human diseases (i.e., methemoglobinemia, non-Hodgkin lymphoma, blue baby syndrome, and even cancer). 2,3Recently, autotrophic denitrification, conducted by chemolithotrophic denitrifiers using inorganic substances, such as H 2 , elemental sulfur (S), S 2 O 3 2− , Fe 0 , Fe 2+ , and iron sulfides, as electron donors has gained intensive attention for efficient NO 3 − removal. 4,5Iron sulfides, such as mackinawite (FeS), pyrite (FeS 2 ), and pyrrhotite (Fe 1−x S, x = 0−0.125),as the highly prevalent sulfide minerals in the Earth's crust and playing an important role in geochemical cycles (nitrogen, phosphorus, sulfur, and iron cycles) in anoxic environments, may represent one of the most promising electron donors for autotrophic denitrification. 5,6n the system of iron sulfides-based autotrophic denitrification (IAD), there are a variety of chemical and biochemical reactions involved in NO 3 − reduction, sulfur oxidation, and iron (Fe 2+ ) oxidation. 7,8Bai et al.'s 7 research found that the NO 3 − chemodenitrification to NH 4 + took place in an iron sulfide (FeS)-based autotrophic denitrification biofilter under neutral conditions and ambient temperature, which might be described by (eq 1): (1) −15 These conflicting results may be due to the use of different iron sulfides, each with its unique structure and properties, thereby not only influencing the autotrophic denitrification efficiency but also affecting the abiotic chemical reduction of NO 3 − in the IAD systems.Previous studies have emphasized the importance of mineral properties in determining autotrophic denitrification performance. 10Therefore, investigating the chemical transformation of NO 3 − with different types of iron sulfides is necessary for an in-depth understanding of potential chemical conversion of NO 3 − by iron sulfides and the cycles of S, Fe, and N in IAD biofilters when chemical and biological NO 3 − transformations take place simultaneously.Several geochemical studies on the N cycle and early life on earth have observed NO 3 − reduction to NH 4 + by FeS under the conditions of high temperature or acidic to neutral pHs.For instance, Wang et al., 16 taking the acid sulfate soil (ASS) environment as the background, explored the N cycle during the NO 3 − chemodenitrification by FeS at ambient temperature (24 °C) and low to neutral pHs (3.5−7).It was found that only 10% of the NO 3 − was reduced to NH 4 + when the pH was 3.5. 16Although low pH is common in natural environments, especially in acidic soils, it is important to study the cycling of S, Fe, and N under relatively mild conditions (15−30 °C, pH 6−7.5, and ambient pressure) simulating conditions commonly found in natural ecosystems and engineered wastewater treatment systems, which is beneficial to maintaining biological reactivity without causing adverse effects on microorganisms or enzymatic processes.Additionally, the optimal pH value of the IAD process is recommended at 6.8−8.2. 17In some other studies in the field of abiogenesis, NH 4 + production through the chemodenitrification of NO 3 − , which was produced in the Hadean oceans by atmospheric reactions possibly driven by electrical discharges or cometary impacts, 18 by iron sulfides, may exist in and around aqueous environments in the Hadean Eon.NH 4 + was considered a fundamental building block for the formation of amino acids and peptides on Hadean Earth, contributing to the origin of life.These research found that only when the temperature was 120 °C, 19 or at low to neutral pH value (4.7−6.9)(eq 2), 20 FeS could reduce NO 3 − to NH 4 + at a low level of 0−6.7% yield: (2) While these studies have investigated the effects of various parameters such as pH and temperature on the chemodenitrification of NO 3 − by FeS, the underlying mechanism of this process remains unclear.Wang et al. 16 detected the presence of reducing sulfur species (such as H 2 S or S 2 O 3 2− ) and speculated that reducing hydrous sulfides may participate in NO 3 − chemodenitrification or passivation inhibiting electron transfer on FeS surface; however, there was no direct evidence.It is yet to be determined whether this reduction process occurs through a homogeneous ion reaction or a heterogeneous solid−liquid reaction.Therefore, a comprehensive study of the chemical reduction mechanisms of iron sulfides and NO 3 − is not only crucial to explore the transformation mechanism of the N, S, and Fe elements in the IAD system but is also of great significance to comprehend the complexity of the N cycle in ecosystems and even explore the origin of life on the Hadean Earth.
Therefore, the purpose of this study was to explore the performance and mechanism of ■ MATERIALS AND METHODS Materials.FeS (fused sticks, Fe 60−67%, S 25%) and FeS 2 (extra pure powder, 1.5−4.5 mm) were purchased from Thermo Fisher Scientific (Geel, Belgium).Pyrrhotite (Fe 56.69%, S 38.46%) 4 was directly obtained from a mine in Tongling City, Anhui Province, China.Before being used in experiments, FeS sticks, FeS 2 powder, and pyrrhotite were pulverized into fine particles (48−550 μm) using a crusher.NaNO 3 served as the NO 3 − −N source, and all solutions were prepared using ultrapure water (18.2MΩ cm).The other reagents employed in this study were of analytical grade.
Experimental Procedures.All experiments were performed in serum glass bottles (160 mL volume) sealed with butyl rubber stoppers.100 mL of NO 3 − −N solution (30 mg/ L) was added into the bottles, as well as a given mass (20 g) of the FeS, FeS 2 , and pyrrhotite, respectively.The dosages of iron sulfide and NO 3 − were representative of typical conditions encountered in IAD biofilters. 7,15,21After flushing with N 2 for 15 min, the bottles were sealed and placed in a shaker at 30 °C; this temperature was based on the recommended temperature (28−32 °C) for autotrophic denitrifiers. 17,22If not specified, the initial pH (6.37) of the experiments was the unadjusted pH of the solutions because it was relatively neutral and close to the recommended pH value (6.8−8.2) for IAD biofilters and autotrophic denitrifiers, 17,22 as well as the pH values commonly found in natural ecosystems (pH 6.5−7.5).After reaction for a set time, 2.0 mL samples were extracted from the bottles and immediately filtered through 0.22 μm nylon syringe filters for the measurement of NO , is all displayed in Text S1.Additionally, the characterization methods on Tafel scans, X-ray fluorescence (XRF), X-ray diffraction (XRD), Xray photoelectron spectroscopy (XPS), field emission scanning electron microscopy (FESEM), and Zetasizer are also shown in Text S1.Although conditions of this study (30 °C, pH 6.37) were largely similar to the previous studies (5−35 °C, pH 5.5− 7.0) 25,27 and all involved Fe(II), previous studies were mainly focused on NO 2 − chemodenitrification or NO 3 − chemodenitrification in the presence of buffer media and catalysts, 28 while in this study, no any buffer media and catalysts were added and the production of NO 2 − was not detected.In Kalina's research, 26 the structure of iron (Fe-rich smectite clay mineral nontronite and the mixed Fe(II)−Fe(III) oxyhydroxide phase green rust) played a crucial role in the process of chemodenitrification of NO 2 − to produce N 2 O, 26 but this iron(II) structure was not detected in the XRD pattern of FeS samples before and after NO 3 − chemodenitrification in this study.Therefore, different NO 3 − chemodenitrification pathways might have resulted in different N 2 O emissions.
During the NO 3 − removal process, the concentrations of aqueous Fe 2+ in the FeS system were about 9.74 mg/L at 24 h and then consistently remained below 0.1 mg/L, and dissolved Fe 3+ was not detectable.These results could be attributed to the pH of the solution increasing from 6.37 to 9.78 in the FeS system during the NO 3 − reduction, leading to the precipitation of iron ions.The concentrations of Fe 2+ in the FeS 2 system were from 66.60 mg/L at 24 h to 2.94 mg/L at 120 h and were consistently around 400 mg/L in the pyrrhotite system (Figure 1c).In comparison to the FeS system, higher concentrations of Fe 2+ were observed in both the FeS 2 and the pyrrhotite systems.This discrepancy may arise from the delayed formation of iron ion precipitation, likely attributable to the pH levels of the FeS 2 −NO 3 − system (5.67) and the pyrrhotite−NO 3 − system (5.97) after 120 h.Alternatively, these differences could be linked to the unique characteristics and reactivity inherent to FeS 2 and pyrrhotite.FeS 2 has a cubic crystal structure, and the presence of sulfur−sulfur bonds within the crystal lattice contributes to its stability. 29The crystal structure of pyrrhotite is less ordered and may contain vacancies or substitutions. 30This structural variability could

Environmental Science & Technology
increase its reactivity compared to FeS 2 potentially leading to a significant dissolution of Fe 2+ .However, no chemical reduction of NO 3 − occurred in these two systems, indicating that Fe 2+ released from iron sulfides in the aqueous phase cannot chemically reduce NO 3 − .The solution after NO 3 − chemodenitrification by FeS was alkaline mainly because FeS is dissolved in a non-oxidizing manner, releasing iron ions into the solution, and the resulting sulfide ions consume protons to form HS − or H 2 S (depending on the pH value). 31In addition, as shown in eq 1, the formation of NH 4 + consumes protons and consequently causes pH to rise.Although the alkaline environment was the result of NO 3 − chemodenitrification by FeS, it was not the decisive factor in promoting NO 3 − chemodenitrification by FeS.When the initial pH was acidic, as discussed in the next section, NO 3 − chemodenitrification by FeS still took place.
In addition, the concentrations of SO 4 2− increased in all three systems (Figure 1d).This may be due to the dissolution of SO 4 2− on the surface of the oxidized iron sulfides into the water or the product SO 4 2− formed after the reaction of the sulfur in the iron sulfides with NO 3 − .However, for the FeS system with NO 3 − reduction occurring, the measured concentrations of SO 4 2− (122.85 ± 6.71 mg/L) were lower than the stoichiometric value (168.53 mg/L) of SO 4 2− calculated according to stoichiometric eqs 1 and 3 7 in consideration that 2.21 mM of NO 3 − −N was reduced to 1.75 mM of NH 4 + −N and 0.18 mM N 2 .Furthermore, the concentrations of S 2 O 3 2− and SO 3 2− in the solution detected by IC were negligible.These results indicate that incomplete sulfur oxidation or electron transfer occurred on the surface of FeS under this circumstance. (3) Impacts of Conditions on NO 3 − Chemodenitrification.The type of nitrates or cations affects the solution properties such as ionic strength and consequently might affect the efficiency and selectivity of chemodenitrification of NO 3 − .Additionally, reaction parameters, such as the mass ratio of FeS/N, reaction temperature, and initial solution pH, might affect the mass transfer, activation, and adsorption during the removal of NO 3 − by FeS.Therefore, it is imperative to examine their impacts on the NO 3 − chemodenitrification efficacy in the FeS system.
It can be seen from Figures 1a and S4 that when three types of NO 3 − (NaNO 3 , KNO 3 , and NH 4 NO 3 ) were used, the chemodenitrification efficiency of NO 3 − by FeS 2 and pyrrhotite was almost negligible, but only occurred by FeS.The chemodenitrification efficiency of KNO 3 , NH 4 NO 3 , and NaNO 3 by FeS (with or without Ca 2+ ) and the selectivity of NH 4 + production did not change greatly.These results indicate that the type of nitrate and the cations would not significantly impact the NO 3 − chemodenitrification.In Figure 2a, it is evident that the NO 3 − chemodenitrification efficacy was obviously improved with increasing the mass ratio of FeS/N Environmental Science & Technology from 1667 to 13 333 because more FeS can provide more active sites.The selectivity of NH 4 + did not change significantly (Figure S2a).Similarly, raising the reaction temperature also accelerated the reduction of NO 3 − (Figure 2b).This was likely because of the accelerated molecular movement accompanying the temperature increase, thereby promoting mass transfer in the heterogeneous system.Through the plots of ln k and 1/T according to the Arrhenius equation (eq S2) (Figure S3), the activation energy (Ea) for NO 3 − −N removal was determined to be 36.78kJ mol −1 .This value exceeds the activation energy typically associated with the diffusion-controlled processes (<20 kJ mol −1 ) and lower than chemical reaction steps (>80 kJ mol −1 ), such as bond breaking, indicating that the NO 3 − chemodenitrification process was dominated by heterogeneous reactions occurring on the FeS surface. 32Furthermore, as the reaction temperature was elevated from 20 to 50 °C, the selectivity of the product NH 4 + increased correspondingly (Figure S2b).This could be attributed to the higher temperature leading to a greater number of NO 3 − molecule bound to the surface of FeS, thereby facilitating the surface-mediated activation of NO 3 − adsorption and NH 4 + generation. 19s illustrated in Figure 2c, the removal efficiency of NO 3 − − N gradually declined from 100% to 70% as the initial pH increased from 2.04 to 11.95.These observations could be clarified by considering the consumption of proton (H + ) ions during the NO 3 − removal reaction, as shown in eq 2. 20 Furthermore, the solution pH has the potential to modify the electrical properties of the FeS surface, influencing its electrostatic interaction (adsorption or repulsion) with NO 3 − . 33Figure 2d displays that the zeta potential values of FeS declined with increasing pH values.When the pH was below pH zpc (pH at the zero point of charge), the positively charged FeS surface could electrostatically attract negatively charged NO 3 − species, facilitating the reductive conversion of NO 3 − .However, an increase in pH would lead to electrostatic exclusion between negatively charged FeS and anionic NO 3 − species, thereby not favoring the reduction efficiency.Figure 2c shows that the reaction rate constants at pH 6.37, 7.34, and 9.48 exhibited slight differences, despite the fact that the proton concentration increased significantly.This could be attributed to the negatively charged surface of FeS when the solution pHs were 6.37, 7.34, and 9.48 (Figure 2d), so there might exist an effect of electrostatic repulsion with the reactant NO 3 − , which resulted in no significant increase in the reaction rate.In addition, even though an increase in the proton concentration might drive the equilibrium toward the side of the reaction that consumed protons, the equilibrium might not shift significantly due to factors like Le Chatelier's principle, which predicts that a system will adjust to counteract changes in conditions (like pH).As observed in Figure S2c, the selectivity of the NH 4 + product during the NO 3 − reduction increased obviously at an initial pH 2.0.This increase could be attributed to the existence of additional positive charges on the surface of FeS (as shown in Figure 2d) at pH 2.04, which could attract more electrostatically negative NO 3 − species.Additionally, the increased concentration of H + in the solution, with H + actively participating in the reaction through the formation of H ads on the FeS surface, increased the probability of a collision with N intermediates.This resulted in a more readily reduction to NH 4 + . 34It is noted that some experimental groups with high reaction rates displayed large standard deviations.NO 3 − chemodenitrification by FeS was a heterogeneous reaction, so slight differences in the distribution or activity of reaction sites in the parallel experiments would cause observed differences in reaction rates.
Characterization of Iron Sulfides.SEM imaging (Figure S5) shows that FeS, FeS 2 , and pyrrhotite pulverized by the crusher were all made of monomer particles and tiny particles.The chemical NO 3 − reduction by FeS was not due to its particle morphology being different from that of FeS 2 and pyrrhotite.XRF (Table S1) analysis shows the main element contents in iron sulfides that had not been exposed to NO 3 − , included 27.60% S and 71.50% Fe in FeS; 39.90% S and 45.40% Fe in FeS 2 ; 34.90% S and 62.60% Fe in pyrrhotite; and other impurities such as Si, Ca, Mg, and Al.The XRF results show that the mass ratios of sulfur and iron in FeS were 27.60% and 71.50%, respectively, which were inconsistent with the theoretical mass ratios of the molecular formula of FeS (sulfur: 36.36% and iron: 63.64%), indicating that excess Fe existed in FeS, compared to S. So, the FeS contained iron oxides or impurity iron power.Based on the actual measured mass content of the sulfur element (27.60%) in the FeS used in this study, it is calculated that the theoretical mass content of the element iron in FeS amounted to 48.3%.This implies that there may exist at maximum 23.2% impurity iron powder in FeS samples, indicating that the maximum amount of impurity iron powder contained in the system of the NO 3 − reduction by 200 g/L FeS was 46 g/L.In order to explore whether 46 g/L of iron powder could contribute to NO 3 − chemodenitrification, the experiment on NO 3 − chemodenitrification by a mixture of 154 g/L FeS and 46 g/L iron powder, only 46 g/L iron powder, 154 g/L FeS, or 200 g/L FeS were conducted.As shown in Figure S6, the efficiency of NO 3 − chemodenitrification by 46 g/L iron powder alone was 22.7% ± 7.5%; The efficiency of NO 3 − chemodenitrification by the mixture of 154 g/L FeS and 46 g/L iron powder (89.4% ± 2.0%) was not significantly improved compared to by 154 g/L FeS alone (88.6% ± 2.5%), and was lower than that of by 200 g/L FeS (97.5% ± 0.1%).Therefore, it can be inferred that the performance of NO 3 − chemodenitrification by FeS was not attributed to the possible presence of iron powder in the FeS and that the NO 3 − chemodenitrification efficiency was not significantly improved when iron powder and FeS coexisted.
To unravel the reasons for the superior NO 3 − removal performance by FeS, Tafel scans were first used to check the electron release ability of FeS, FeS 2 , and pyrrhotite in solutions from thermodynamic point of view by measuring the free corrosion potential, 35 as the more negative value of the free corrosion potential reflects that the material is more likely to lose electrons. 36The free corrosion potentials of FeS, FeS 2 , and pyrrhotite amounted to −0.16, −0.09, and −0.06 V, respectively (Figure 3), suggesting that the electron release from FeS was easier than that from FeS 2 and pyrrhotite.
XRD spectra clearly indicate that FeS, FeS 2 , and pyrrhotite were highly pure crystal structures (Figure 4).For FeS, the characteristic diffraction peaks at 2θ of 29.92, 33.67, 43.15, and 53.12 Å were assigned to the phases (110), ( 112), (114), and (300) for troilite−2H (FeS), respectively (PDF no.37−0477).The characteristic diffraction peaks at 2θ of 36.04,41.93, and 60.76 Å were assigned to the phases (111), (200), and (220) for wustite and syn (FeO), respectively (PDF no.06−0615).Reacted solids after exposure to NO 3 − showed a significant reduction in the characteristic peak of troilite−2H (FeS) (Figure 4a).No characteristic peaks of iron powder were found in the XRD characterization results of FeS before and after the Environmental Science & Technology reaction, which also indicates that there was no elemental iron or the trace amount of elemental iron was below the limit of detection in FeS.For FeS 2 , the diffraction peaks observed at 2θ values of 28.51, 33.08, 37.11, 40.78, 47.41, and 56.28 Å were attributed to the (111), ( 200), ( 210), ( 211), (220), and (311) crystal planes of pyrite (FeS 2 ), respectively (PDF no.42− 1340). 37For pyrrhotite, the peaks at 2θ of 29.92, 33.84, 43.76, and 53.11Å were in accordance with ( 200), ( 205), (2010), and (220) crystal planes of pyrrhotite−5T (Fe 1−x S), respectively (PDF no.29−0724). 37The characteristic diffraction peaks at 2θ of 44.03 and 71.55 Å were assigned to the phases ( 206) and ( 406) for pyrrhotite−3T and syn (Fe 7 S 8 ), respectively (PDF no.24−0220), and the characteristic diffraction peak at 2θ of 65.44 Å was assigned to the phases (800) for pyrrhotite−4 M (Fe 7 S 8 ) (PDF no.29−0723).After exposure to NO 3 − , the characteristic peaks of FeS 2 and pyrrhotite solids did not change significantly compared to those of FeS (Figure 4).The differences in the XRD spectra of the three iron sulfides confirmed the chemical reduction reaction between NO 3 − and FeS.XPS was used to analyze the changes in surface compounds and electronic states of FeS, FeS 2 , and pyrrhotite before and after NO 3 − reduction.As depicted in Figure 5a, the XPS spectra of Fe 2p peaks of FeS at 710.61, 712.16/718.20/725.85, and 724.08/731.99 eV were identified as corresponding to Fe(II)−S, Fe(III)−O, and Fe(II)−O, respectively, indicating the partial oxidation of Fe(II) on the FeS surface. 38fter the NO 3 − reduction reaction, the percentage of Fe(II) species (Fe(II)−S and Fe(II)−O) decreased from 27.88% to 14.69%, indicating the oxidation of Fe(II) into Fe(III) species by NO 3 − on the FeS surface.In contrast, the Fe(II) species on the FeS 2 and pyrrhotite surface did not decrease significantly before and after the reaction (Table S2 and Figure 5b,c).Furthermore, for the S species on the FeS surface, S(−II), S 2 (−II), and S n (−II) (n = 3, 4, ...) comprised approximately 49.64% of the total S in the fresh FeS sample (Figure 5d); after the reaction with NO 3 − , the disappearance of S 2 (−II) (16.43%) in the S 2p and the decrease of S(−II) (from 15.03% to 7.84%), together with the increase in the proportion of S n (−II) (from 18.18% to 32.94%) and S(VI) (from 50.35% to 59.22%), indicate that these S species might serve as the reductants for NO 3 − removal and/or the electron donor to promote the Fe(II)/Fe(III) cycle.In contrast, for FeS 2 and pyrrhotite, the concentrations of S(VI) decreased significantly (Table S2 and Figure 5e,f).Considering the obvious increase of the SO 4 2− concentrations in solutions (Figure 1d), inherent S(VI) might be detached from the surface of FeS 2 and pyrrhotite during the reaction.To conclude, both XRD and XPS results confirm that FeS participated in the chemical reduction of NO 3 − , while FeS 2 and pyrrhotite did not.

Role of Aqueous Fe and S Species in the Process of NO 3
− Chemodenitrification on FeS.Both Fe(II) present on the FeS surface and Fe 2+ released from the dissolution of FeS can serve as the active species for NO 3 − chemodenitrification.However, as shown in Figure 6b, soluble Fe 2+ did not have the ability to reduce the level of NO 3 − (30 °C, without pH adjustment).In addition, as shown in Figure 1c, the concentrations of aqueous Fe 2+ were quite low.Moreover, the XPS spectra for the Fe 2p regions of pristine and used FeS show that the proportion of Fe(II) obviously decreased during the reduction of NO 3 − .Therefore, soluble Fe 2+ was not the main cause for the chemical reduction of NO 3 − .As the S 2− released from FeS undergoes hydrolysis and electron donation processes, it might be transformed into various sulfur species such as HSO 3 − , HS − , H 2 S, S 2 O 3 2− , and SO 3 2− . 39,40The reduction of NO 3 − by some reducing sulfur species, such as H 2 S, is thermodynamically favorable over a wide pH range, 41 so the contributions of sulfur species in the NO 3 − reduction process need to be considered.According to  , and HSO 3 − systems was almost negligible (Figure 6b).Therefore, it is speculated that the sulfur species in the aqueous phase would not be the main active species for NO 3 − reduction.
To assess the role of •H in the NO 3 − chemodenitrification on FeS, experiments were conducted with tertiary butanol (t-BuOH), a scavenger of •H that converts it into inert 2-methyl-2-propanol radicals. 43As depicted in Figure 7a, the efficiency

Environmental Science & Technology
of NO 3 − chemodenitrification decreased from 97.5% to 84.2% with t-BuOH (0.1 M) and then stabilized with increasing t-BuOH dosage.This suggests that approximately 13% of NO 3 − chemodenitrification might be attributed to surface-adsorbed •H.Consequently, it can be inferred that the majority of the NO 3 − reduction was not mediated by reducing •H but by direct electron transfer coupled with H ads formation facilitated by H + on the FeS surface.To further explore the NO 3 − chemodenitrification mechanism at the active sites on the FeS surface, 2,2'−bipyridyl (BPY), capable of chelating Fe(II) to impede electron transfer to the oxidants on the particle surface, 44 was introduced into the reaction system.As anticipated, the chemodenitrification efficiency of NO 3 − −N by FeS decreased from 97.5% to 66.5% in the presence of BPY (2.0 mM) and then decreased to 15.1% as the BPY dosage was increased to 32 mM (Figure 7b), confirming that Fe(II) was the main active site on FeS for NO 3 − chemodenitrification. From the above results, high concentrations of sulfur vacancies existed in the FeS, and it provided more active sites for NO 3 − reduction.The vacancies can also increase the adsorption capacity of FeS for NO 3 − , allowing for the more efficient reduction of NO 3 − to NH 4 + .DFT calculations were further employed to examine the adsorption behavior of the NO 3 − and H 2 O molecules on the sulfur vacancies on the FeS surface.Following the optimization of the most stable adsorption modes, the calculated adsorption energy (E ads ) of O atoms in the NO 3 − and H 2 O molecule on the sulfur vacancy of FeS surface was −1.96 eV and −0.82 eV (Table S3).This value was much lower than that of the N atom (−0.79 eV) in the NO 3 − molecule and the H atom (−0.19 eV) in the H 2 O molecule on the FeS surface with sulfur vacancy available (Table S3), suggesting favorable adsorption of the O atoms of the NO 3 − and H 2 O molecule onto the sulfur vacancy-rich FeS.Furthermore, the bond dissociation energy (D 0 ) for the O atom and N atom within the NO 3 − molecule on the sulfur vacancy of the FeS surface was determined to be 0.21 and 1.03 eV, respectively (Table S3), indicating that the other O atoms in the NO 3 − were preferentially dissociated on the sulfur vacancies.Similarly, the D 0 values of the O atom and the H atom in the H 2 O molecule on the sulfur vacancy of the FeS surface were 0.75 and 0.40 eV, respectively (Table S3), and implied that H atoms in H 2 O molecule were more prone to dissociation, forming H ads , which interacted with the deoxygenated NO 3 − .This sequence of deoxygenation and hydrogenation processes actively promoted the formation of NH 4 + .Based on the above results and systematic discussion, the possible reduction pathways of the NO 3 − −FeS system is proposed and shown in Figure 8.Approximately 13% of the NO 3 − reduction (Figure 7a) is related to the role of •H generated by sulfur vacancies on the FeS surface and H 2 O molecules.Most of the NO 3 − reduction is dominated by direct electron transfer and H ads .In general, the O atoms of NO 3 − tend to be adsorbed on the sulfur vacancy sites.At the same time, the O atoms in water molecules are adsorbed on the sulfur vacancies, followed by water dissociation, which generates H ads on the nearby sulfur vacancy sites.Then, Nintermediates adsorbed on the sulfur vacancies of FeS are converted to NH 4 + by electron transfer coupled with hydrogenation, and a small number of N-intermediates are converted to N 2 .

■ ENVIRONMENTAL IMPLICATIONS
Three kinds of iron sulfides, such as FeS, FeS 2 , and pyrrhotite, were studied for the NO 3 − chemodenitrification under mild conditions.Chemical NO 3 − reduction by FeS occurred but not by FeS 2 and pyrrhotite.FeS had a high concentration of sulfur vacancies and quickly released electrons, which may be the main reason for its efficient chemical reduction of NO 3 − .Experimental results and DFT calculations demonstrate that Fe(II) on the FeS surface was the primary reactive site for NO 3 − chemodenitrification and that FeS possessing sulfur  − molecule and promoted intrinsic activity for H ads formation through H 2 O dissociation, thus leading to a heightened selectivity in NH 4 + formation (82.6%).Although NH 4 + is considered an important indicator to nitrogen pollution in wastewater and water environment, NH 4 + /NH 3 is a type of nitrogen fertilizer and even can be used as a fuel. 45,46Therefore, the highly selective NH 4 + production by means of chemodenitrification of NO 3 − allows us to envision that NO 3 − , a pollutant, might be recovered as a fertilizer and fuel through this chemodenitrification process.However, comprehensive research on real wastewater should be conducted in the future.In addition, this research helps in improving our understanding of the basic science behind chemical NO 3 − reduction and, in turn, advancing the NO 3 − remediation by iron sulfides-based technology.Furthermore, this work also provides broader insights into the field of geochemistry for understanding the complexity of N, Fe, and S cycles in ecosystems and even the conditions for the emergence of life on early Earth.However, extrapolating these results to a broader context requires careful consideration of the limitations inherent in laboratory experiments and the complexity of the Hadean Earth environment.
NO 3 − chemodenitrification by different iron sulfides.The specific objectives of this study were (1) to assess NO 3 − chemical reduction by three common iron sulfides: FeS, ferrous disulfide (FeS 2 ), and pyrrhotite; (2) to investigate the chemodenitrification efficiency of NO 3 − with FeS based on the effects of type of nitrates, m(FeS)/ m(NO 3 − −N) ratio, reaction temperature, and initial pH values; and (3) to explore the mechanism of NO 3 − chemodenitrification by FeS through the identification of reactive species, quenching experiments, and density functional theory (DFT) calculations.
Sulfides.Figure 1a presents NO 3− chemodenitrification by iron sulfides under anoxic conditions.FeS chemically reduced NO 3 − at 30 °C and unadjusted initial pH, and the NO 3 − −N reduction efficiency was 97.5% ± 0.1% after 120 h, while the NO 3 − chemodenitrification was almost negligible within 120 h when FeS 2 and pyrrhotite were the reactants.It was observed that the main NO 3 − chemodenitrification product by FeS was NH 4 + , and the concentration of NH 4 + −N increased to 24.53 (±1.95) mg/L at 120 h.NO 2 − was not detected during the reduction of NO 3− by FeS (FigureS1).However, the amount of NH 4 + −N produced (24.53 ± 1.95 mg/L) did not account for the total NO 3 − −N removed (29.70 ± 0.81 mg/L) from solution by FeS (Figure1a,b); the other product of NO 3 − reduction was N 2 with the concentrations of NO, NO 2 , and N 2 O in the headspace gas phase below the limit of detection.Many previous studies observed N 2 O emission during the chemodenitrification process of Fe(II) and NO 3 − /NO 2 − . 23−27 However, in the present study, negligible N 2 O was produced.

Figure 2 .
Figure 2. Effects of reaction conditions on NO 3 − −N removal for FeS: (a) Mass ratio of FeS/N, (b) temperature, and (c) pH value.Except for the investigated parameters, other parameters were fixed: [FeS] 0 = 200 g/L, [NO 3 − −N] 0 = 30 mg/L, and T = 30 °C, without pH adjustment.(d) Zeta potential analysis of FeS in relation to the pH of the solution.

Figure 4 .
Figure 4. XRD patterns of (a) FeS, (b) FeS 2 , and (c) pyrrhotite samples before (pristine) and after (used) the reaction compared with the standard materials.

NO 3 −
Chemodenitrification Pathway by FeS.From the above analysis, it is quite clear that the reducing ions (Fe 2+ , HS − , HSO 3 − , SO 3 2− , and S 2 O 3 2− ) in the solution were not the main active species for reducing NO 3 − .Therefore, NO 3 − reduction was dominated by the interface processes on the FeS particle surface rather than the soluble active species leached from the material.The XPS peak showed the atomic ratios of S:Fe were 0.54, 1.93, and 1.06 in FeS, FeS 2 , and pyrrhotite, respectively.This indicates that high concentrations of sulfur vacancies (Vs) existed on FeS.At the sulfur vacancies sites, the back-donation of localized electrons would enhance the heterolytic dissociation of adsorbed H 2 O molecules, leading to the generation of •H or (e − and H + ) (eqs 4

Figure 8 .
Figure 8. Schematic illustration of the NO 3 − chemodenitrification mechanism in the FeS system.
Analytical methods; density functional theory (DFT) calculations; variation of NO 2 − −N during the reduction of NO 3 − by iron sulfides; effects of operation parameters on the ratios of NH 4 + production vs NO 3 − reduction for FeS; plot of ln k and 1/T of the NO 3 − reduction; effects of different types of nitrate and cations on NO 3 − removal and the ratios of NH 4 + production vs NO 3 − reduction by iron sulfides; SEM images, element components, and XPS results based on curve fitting for Fe 2p and S 2p peaks of FeS, FeS 2 , and pyrrhotite; variation of NO 3 − −N during the reduction of NO 3 − by iron power, FeS, and a mixture of iron power and FeS; adsorption energies (E ads ) and bond dissociation energy (D 0 ) of the atoms in the H 2 O and NO 3 − molecules on the sulfur vacancy on the FeS (114) surfaces (PDF) ■ AUTHOR INFORMATION Corresponding Author Xinmin Zhan − Civil Engineering, School of Engineering, College of Science and Engineering, University of Galway, Galway H91 TK33, Ireland; orcid.org/0000-0002-9101-1404;Phone: + 353−08−7708−0629; Email: xinmin.zhan@universityofgalway.ie